Or SNaval Research LaboratoryWashington, DC 20375-5000

NRL Memorandum Report 6893

AD-A242 122

A Pulsed Nitrogen Laser for Optical Plasma Diagnostics

A. W. DESILVA*

SFA, Inc.Landover, Maryland

The Universin' of MarylandCollege Park, Maryland

and

J. D. SETHIAN

Charged Particle Physics BranchPlasma Physics Division

October 1, 1991

91-15198

Approved for public release. distribution unlimited.

11 4

r Form ApprovedREPORT DOCUMENTATION PAGE O MB No 0704-0188

PuO1,C ten~fo r en t0, MhS co11eCTIOn of mtortmatson s estimnated to a-erage 'iiU? oer esoorse ircudtng the tite tor reuie-iq "structions searchihq e.'5t-n9 data sourcesgathering and mtaintaining the data needed, and comhoretsig and re-ewnq the ,11peXton o1 itormationI Send commtent% regarding this ou.,den estimate or anv other aspect oft thiscoflectiOn 0oaf t'mtatiOA ncltudifg suggesions 'or reducing tINS ourdeh to AaSthngton rueadauarters Servces Direct orate for information Opierations and Redonit. 12 15ceftrsonDa,iSrtigiha Y, S uite 2 04 Ariingtott. 4A 22202-4301 and to the OWite of management id Budget Paperwsorr Rieduction Project (0 C4.0188). Washington (IC 20503

1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED

1 1991 October 14. TITLE AND SUBTITLE 5. FUNDING NUMBERS

A Pulsed Nitrogen Laser for Optical Plasma Diagnostics

6. AUTHOR(S) 47-1894-01Alan W. DeSilva* and John D. Sethian

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION

Plasma Physics Division REPORT NUMBERNaval Research Laboratory4555 Overlook Avenue NRL MemorandumWashington, D.C. Report 6893

9. SPONSORING/ MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSORING /MONITORINGAGENCY REPORT NUMBER

ONRArlington, Va 22217-5000

11. SUPPLEMENTARY NOTES

*SFA, Inc.. Landover. Maryland andThe University of Maryland. College Park. Maryland

12a. DISTRIBUTION / AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE

Approved for public release, distribution unlimited

13. ABSTRACT (Maximum 200 words)

A nitrogen laser oscillator-amplifier system has been constructed for use as a fast pulsedlight source for schlieren, shadowgraph, Moird&Schlieren, or interferometric diagnostics ofpulsed plasmas. It utilizes an atmospheric pressure nitrogen laser as an oscillator and alow pressure transverse discharge as an amplifier. The output beam has a rectangularcross section with dimensions 5mm x 15mm, a pulse length of about 1 nsec, and is ofsuffcient optical quality that it may be used for interferometry. The beam energy is about

is 1 mJoule per pulse.

14 SUBJECT TERMS 15. NUMBER 01~ PAGES

Plasma Diagnostic, Pulsed Nitrogen laser, Interferi:meter, 162RC 0CDShadowgraph, Moir6-Schlieren 1.PIECD

17, SECURITY CLASSIFICATION 18 SECURITY CLASSIFICATION 19 SECURITY CLASSIFICATION 20. LIMITATION OF ABSTRACT

OF REPORT OF THIS PAGE OF ABSTRACT

Unclassified Unclassified Unclassified SAR

..................................................... . 9 H

TABLE OF CONTENTS

I. Introduction 1

II. Oscillator 1

In. Amplifier 2

IV. Optics 3

V. Interferometer 4

VI. Moird Schlieren 6

VII. Summary & Acknowledgements 8

VII. References 8

'* k.I /

" -" *

" "' vtr

iii

A PULSED NITROGEN LASER FOR OPTICAL PLASMA DIAGNOSTICS

A.W. DeSilvaSFA, Inc and The University of Maryland

-and-J.D. Sethian

Plasma Physics DivisionNaval Research Laboratory

I. INTRODUCTION

A pulsed laser light source was needed for diagnostics of the dense Z-pinch plasmasproduced by the NRL ZFX experiment1 . The laser was required to have a pulse length ofless than 1 nsec in order to freeze the rapidly changing conditions in the pinch, acoherence length suitable for interferometry, and sufficient power to overcome the intensecontinuum radiation from the plasma.

The required short pulse can be achieved with the typical atmospheric pressuretransverse discharge nitrogen laser, but the beam has poor optical quality2 . A lowpressure nitrogen laser has the necessary optical quality but at the expense of anincreased pulse width (at least several nsec) 3 ,4 . While it is possible to build anatmospheric pressure, short pulse laser of interferometric quality5 , a more efficientmethod is to use an oscillator-amplifier system2,6,7,8,9,10,11. This is the approachdescribed in this paper. The beam is generated by an atmospheric pressure transversedischarge laser, improved by spatial filtering, and then amplified by a single pass lowpressure transverse discharge amplifier.

II. OSCILLATOR

The oscillator was built by F. Sandel, and utilizes a Blumlein pulse lineconstructed from 0.7 mm thick fiberglass printed circuit board material. The constructionis pictured in Figure 3. The line is split on the upper side, and is an unbroken conductor(represented in the figure by a dotted line) on the opposite side. The electrodes arealuminum 9.2 mm thick, fully radiused, and pressed against the printed circuit board by alarger lucite plate above. They are separated to form a discharge gap about 4 mm wide.Adjustment screws make it possible to finely adjust the plates for parallelism. Thisadjustment, once set, rarely needs to be disturbed. It is made by watching the dischargefrom above and setting the plates to obtain the most uniform discharge. The nitrogenworking gas (Ultra High Purity Grade, 99.995% pure) is delivered straight from the bottlethrough a hole in the upper lucite cover and flows freely out both ends, displacingatmospheric air in the process. A flow rate of 1.0 SCFH is used, but is not critical.

Manucrpt appro'ed August 29. 1991

The oscillator is switched by a spark gap having 12 mm diameter brass electrodeswith a midplane tungsten needle held at half the potential by 100 MO resistors. In orderto assure the switch can be triggered with low jitter, it must be run at a potential justbelow self-break (-14 kV, which is "32" on the power supply dial), and the needle must besharp. If the needle is removed for sharpening, it will be necessary to reset the electrodesafterwards, since the needle rarely returns to its previous position. The gap should beabout 2-2.5 mm wide, with the needle centered, and the gap pressurized with nitrogen to10 PS1G. Both electrodes have screwdriver adjustments to allow this. The voltageshould not be allowed to run above 15 kV, as this is near the breakdown for the fiberglassinsulation of the Blumlein line. (If a breakdown does occur the resulting damagedinsulator can be repaired with RTV.)

The oscillator is fitted with a mirror on one side of the discharge cavity, whicheffectively doubles the length of the cavity. This mirror rarely needs adjustment. If itdoes, however, it is sufficient to simply look along the channel and align the reflection ofthe discharge electrodes with the electrodes themselves. We have also tried aligning themirror by watching the output laser pulse, but neither method seems to offer greatadvantages. The emitted beam is more divergent in the horizontal plane (parallel tocurrent flow) than in the vertical. This is obvious by just looking at the beam at pointsfarther from the laser, or by observing the burn spot made at the focal plane of the 50mmlens. A photodiode is oriented to look at a small fraction of the beam tapped off by amicroscope slide. The output of this photodiode is positive and about 10 volts and may beused as synchronization pulse. Because the output is very narrow in time, it is useful tointegrate it with a 50-200 nsec integrator to make it more visible.

IH. AMPLIFIER

The amplifier (Figs 2a,b) design is similar to the oscillator described by Fisher et.al.3 ,4 . Two rods 2.54 cm in diameter and 45.7 cm long serve as the electrodes. Theirfacing sides have been planed off to produce a flat about 2 cm. wide, and the edges arerounded to approximate a Bruce profile12 Moir6-Schlieren. The electrode gap is 2 cm.The electrodes are mounted in a lucite tube that is 5 cm id and 61 cm lon! %nd fitted with2.54 cm diameter high quality quartz windows. These windows ar. :anted about 2"from the axis perpendicular to prevent the amplifier from oscilla .ing spontaneously.(Somewhat surprisingly, using anti-reflection coated windows alone was insufficient.)

Electrodes of both brass and carbon have been tried. The brass electrodesproduced highly non-uniform discharges even with a cnrting of aquadag. With thecarbon electrodes it is possible to find a range of press ures and operating voltage thatresults in many small arcs that merge to a sufficienfiy uniform discharge. Two sets of

carbon electrodes have been tried. One set had nearly flat facing surfaces with only

2

slightly rounded edges, and produced a discharge that uniformly filled the 18 mm wideregion between the electrodes. T" As amplified that was almost as wide as the electrodewidth. The other set had a smaller radius of curvature, produced a 5 mm diameterdischarge, and amplified a correspondingly tighter beam.

The amplifier is electrically driven by a parallel-plate pulse line having a width of44 cm (i.e. almost equal to the electrode length), as shown in Figure 2b. Each side of theline has a characteristic impedance, Zo - 1.6 0, and is 1.2 meters long. It is doublyfolded to make a compact structure that extends only about 30 cm to each side of thedischarge tube. The dielectric is several layers of Mylar with a total thickness of 1.6mm.This line is pulse-charged with a quarter period of 80 nsec from a bank of eleven 2700 pf @40 kV doorknob capacitors. The discharge tube (i.e. laser) self-breaks with about a 10nsec jitter. The peak amplifier current is about 20kA, so the current density for thesmaller radius electrodes is about 900 Amp/cm2 and for the flatter electrodes about afactor of three smaller. As expected, the amplified light beam is distinctly less intense inthe latter case.

Because the osciilator output must be synchronized with the peak amplifierdischarge current to within about ± 2 nsec it is necessary to provide a precise trigger.This is achieved with the electrical circuit shown schematically in Figure 3a, and depictedin Fig 3b. An external trigger fires the amplifier switch which simultaneously startsthe 80 nsec long charging cyci, of the pulse line and sends a pulse down two 80 nsec longRG/8 cables which are connected across the amplifier switch output. One of these cablesfires a sparker placed at one end of the amplifier just as the amplifier voltage is reachingmaximum, causing the amplifier discharge to fire with about 1 nsec jitter. The secondcable triggers the oscillator switch, which produces its beam within about 1 nsec. Thecables may be pruned to obtain the optimum performance.

The entire N2 laser system is mounted inside an electrically shielded box. Thebox also prevents dust from collecting on the laser components. The four sides of the boxand the top are readily removable with Dzus fasteners to allow access to the laser. Twophotos of the laser system is shown in Figure 4a and 4b.

IV. OPTICS

The optical path is shown schematically in Fig 5. The output of the oscillatorreflects from a 45" dielectric mirror that is nearly 100% reflecting at 337 nm, buttransparent in the visible. At this point, a He-Ne laser beam is introduced to lie along

the N 2 laser path for alignment purposes. The beam(s) are brought into focus by a 50mm focal length quartz lens, and spatially filtered with a 10 p. m diameter pinhole. Thispinhole may be moved by screws along two axes perpendicular to the beam. The

3

adjustment is exceedingly delicate, and it is difficult to find the right spot with the He-Nebeam if it is lost. (A 10 g m hole is not very large, and the focus of the He-Ne laser beamis only about 30 g m in diameter.)

The N2 laser focal spot is larger, measuring about 100 stm x 250 ;rn, but since itmay only be seen when the laser is pulsed, it is usually easier to locate the pinhole usingthe He-Ne beam. There is also an adjustment to place the pinhole in the focal plane ofthe lens. This can be done by carefully observing the size of the burn mark made by theoscillator pulse on a target, such as exposed polaroid film, that is mounted in place of thepinhole. Once so positioned, it should not be necessary to change this setting.

Light passing through the pinhole is collimated by a 175 mm focal length quartzlens, which forms a beam about 2 cm in diameter, and is directed by a mirror through theamplifier. The gain profile of the amplifier discharge is reasonably uniform in thevertical direction, filling the 20 mm space between the electrodes, and about 5 mm wide inthe horizontal direction (in the case of the more curved electrodes). Thus the amplifiedbeam is rectangular in cross-section. With the flatter electrodes the output beam waswider but lacked the intensity required for interferometry.

To check for laser operation it is convenient to place a fluorescent target such asexposed Polaroid film inside the laser cabinet covering the exit port. As the inside of thecabinet is dark, one can easily observe the output while operating the laser. An ill-defined wash of white or bluish light on the screen indicates only that one is seeing thedischarge light. On the other hand, bright well-defined spot indicates the amplifier isworking properly.

If no amplified pulse appears, there are several things to check. First the oscillatorgas should be checked with the flow meter. One may then try turning the oscillatorvoltage up a bit (Not over 15 kV!) and varying the amplifier fill pressure. The amplifierruns with a steady flow of gas through the discharge chamber, so the exhaust valveleading to the pump should be fully open. The pressure should be set to about 60 Torr onthe gauge, but it may be varied in the 40-80 Torr range. If the pressure is wrong, thedischarge tends to arc.

V. INTERFEROMETER

The interferometer (Fig 6) is in a Mach-Zehnder arrangement with the referencebeam displaced only 16 mm to the side of the scene beam. This unconventional featureallows both beams to pass through the 50 mm diameter windows of the ZFX vacuumchamber but only the scene beam to intercept the fiber. The reference beam is off to theside and is undisturbed by the plasma until quite late in the discharge. This setup wasadopted owing to the practical difficulty of maintaining two beam splitters and two

4

mirrors aligned. An incidental advantage is that the optical paths are automaticallycompensated and equalized as both pass through the same windows. The latter isnecessary because of the short (< 1 mm) coherence length of the laser.

Two rigid stands were constructed, with one beamsplitter and one mirror mountedon each stand. These optics may be pre-aligned on the bench and then moved to theexperimental area. The beamsplitters are on quartz substrates with a small wedgeangle so the transmitted beam is slightly deflected. The reflecting surface should beplaced to the outside when orienting the beamsplitters (reflecting surface facing the laseron the input side, and facing the camera on the output side). In this orientation thetransmitted intensity of the two beams is nearly equal.

The two beams emerging from the first beamsplitter/mirror set must pass throughthe 50 mm diameter vacuum windows and the scene beam must be centered in thewindow to illuminate the fiber. The 16 mm diameter beams should be separated by16mm to prevent both overlapping and vignetting by the edges of the vacuum windows.The scene and reference beams are recombined by the second beamsplitter-mirror set andpass on to a camera. A 350 mm focal length lens is placed after the beamsplitter andfocuses the fiber onto the film. By placing a small aperture at the focal point of the lensmuch of the plasma light will be excluded but the laser beam will pass through throughwithout attenuation.

To align the interferometer it is best to first ensure that the scene beam passesthrough the first beamsplitter and then through the center of both of the vacuumwindows. An auxiliary gas laser directed backward through the chamber is helpful.Alignment is easier if a target is installed in place of the fiber. Small adjustments maybe made by the mirror adjustment screws on the last mirror on the laser stand. Rotatingthe entire beamsplitter/mirror assembly will change the separation of the beams slightly,but it will be close enough if the edge of the stand is parallel to a radius of the ZFX loadchamber. The parallelism of the two beams may now be adjusted with the mirroradjustment screws on the input side. The two beams should now pass through the twoholes in the mask on the output side. Getting this setting right is easier if the inputbeam apert ire is decreased using the iris installed on the input side stand.

The trick in alignment at the output side is to make sure that the scene beam,which reflects first from the mirror, hits the beamsplitter in the same spot as thereference beam. This may be adjusted using the mirror adjustment screws. Once thisis set the beams should be observed from as far away as possible, and the beamsplitterscrews adjusted to bring them together. The fringes should then appear readily. Thisis all done without the focussing lens in place. It may then be installed, followed by thecamera and aperture. In order to exclude plasma light we found it necessary to installtwo wide band filters and one narrow band filter in front of the camera. Finally, it is

5

advisable to put a shutter in the beam, to prevent exposure of the film during charging ofZFX in case the laser prefires.

VI MOIRE SCHLIEREN

Moir6-Schlieren interferometry is a technique for measuring the electron densitygradient in a plasma. Its output is an image of the plasma, formed in interference fringes,that appear quite similar to the fringes of ordinary interferometry. The displacement ofthe fringes is proportional to the line-of-sight integral of the transverse (to the line ofsight) electron density gradient in the direction perpendicular to the fringes. Moird-Schlieren was chosen over ordinary interferometry because the beam quailityrequirements are less demanding, the hardware is simpler and cheaper, and the setup iseasier.

The plasma region is illuminated by a collimated beam of monochromatic lightfrom a laser. After passing through the plasma, the beam passes through a pair ofidentical coarse transmission gratings called "Ronchi rulings". These are separated by asmall distance D, and oriented perpendicular to the optic axis, but rotated about the opticaxis so that their rulings depart from parallelism by a small angle a. The beam is thenfocussed onto a screen by a simple lens, and the first order interference maximum allowedto pass through a small aperture in the screen. The recording film is placed farther on,so that it is in an object-image relationship with the plasma region ( Figure 7).

While it is possible to analyse a Moird-Schlieren system by invoking ray optics andtreating the rulings as venetian blinds that simply cast overlapping shadows, thisapproach lacks insight. In particular it does not explain why the first interferencemaximum should be chosen. A more illuminating approach is to think of the variousinterference orders emerging from one of the gratings as being simple plane wavesdisplaced in angle from the laser beam direction by the usual interference angles On =sin'l(n) d), where x is the wavelength, d the grating spacing, and n is an integeridentifying the order. Since the two gratings are rotated with respect to one another bythe angle a, the first order beam from one will not be parallel to that from the other, buthave an angle =ae n between them (for on

gratings at a small angle 6 with respect to the optic axis (Fig. 8). For the first orderbeams, )/d = sin( y)+sin(6). If the angle y has been selected to give an interferencemaximum on a distant screen, then the condition of a maximum is:

AL = nX = D/cosy - D/cos6 + (b + c)sin6, (1)

where:b = D tan6, and c = D tan6. (2)

With 6

Combining Equations 7 and 3, we find the condition for a fringe shift of m fringes is

f(dne/dy) dz = md/c),2D. (8)

As an example, consider the "halo" around a 50 gm diameter fiber and assume that

(dne/dy) - 1020 cM- 3/10-2 cM - 1022 cm -4 .

This gradient will only be felt in a region of about the size of the radius, so

f(dne/dy) dz = 1020 cm -4 .

For a grating with 20 lines/mm, d = 5 x 10 -3 cm, and if it assumed that D = 1 cm, then m= 1. Thus in the vicinity of the fiber the deviation would be about one fringe width.

VII SUMMARY & ACKNOWLEDGEMENTS

The N2 laser oscillator/amplifier system is now ready for deployment. The authorsare grateful for the assistance of Mr. K. A. Gerber and the technical assistance of Mr J.P.Picciotta, Mr. D.R. Evenson, and Mr. K.M. Coffey. This work was supported by the Officeof Naval research and the U.S. department of Energy.

VI1 REFERENCES:

1. J.D. Sethian, A.E. Robson, K.A. Gerber and A.W. DeSilva, in Proceedings of theSecond International Conference on High Density Pinches, edited by N.R. Pereira,J. Davis and N. Rostoker (American Institute of Physics, Conference proceedings195, N.Y. 1989), p 26-28.

2. E. E. Bergmann, Appl. Phys. Lett. 31, 661 (1977).

3. E. Bar-Avraham, A. Fisher, F. Mako, J. Peacock, and J. Shiloh, IEEE Trans. onPlasma Science PS-6, 296 (1978).

4. A. Fisher, J. Peacock and E. K. C. Lee, Rev. Sci. Instrum. 49,395 (1978).

5. H. Salzmann and H. Strohwald, Optics Comm 12, 370 (1974).

8

6. D. M. Rayner, P. A. Hackett, and C. Willis, Rev. Sci.Instrum 52, 1579 (1981).

7. K. Kagawa, M. Tani, N. Shibata, R. Ueno, and M. Ueda, J. Phys. E: Sci. Instrum.15, 1192 (1982).

8. S. Nakamura, H. Hara, K. Umezu, and H. Takuma, Japan J. Appl. Phys. 15, 2491(1976).

9. I. Santa, S. Szatmari, B. Nemet, and J. Hebling, Optics Comm 41, 59 (1982).

10. T. Jitsuno, J. Phys. D.: Appl. Phys., 13, 1405 (1980).

11. F. Docchio, V. Magni, and R. Ramponi, Rev. Sci. Instrum. 55, 477 (1984).

12. F.M. Bruce, J.IEE, 94, 138 (1947).

9

spark gap switch

IN

Blumlein conductor

lumlein conductor

aluminum electrode

PC board lucite

Figure 1: The Nitrogen laser oscillator

10

aluminumground i oflte cneline

"~NNXXNn N %NONNNNNNXXNXX\N"N xN~

almiumseond mylar-finuaed quartzwindow

44 cm

~~~~~Figure 2b T ecinop ieo nitrogen laser amplifier.igtedobyfle

paalelpltelieth dscare ub, ndth eecroes

coax cables

amp switch trigger .-amplifier trigger

amplifier

80nsecoscillator trigger

osc sparkgap I 4==-*-oscillatorswitch

Figure 3a: Complete circuit of nitrogen laser amplifier plus oscillator

2700 pfI11e

charging resistor

AP~E2700 PF

CS-- PIGrRIGGERED SPARK GAP,'capacitor at other DO

ftnd o' cable to 0

Figure 3b: Switching circuit for amplifier.

12

"'OPTICAL PATH

AMPLIFIER

" OSCILLATOR

TRIGGER GENERATOR:

H V SUPPLY -Zi1

H V SUPPLY

Figure 4b: Photo of N2 laser system, top perspective showing the support structure,electronic components and relative positions of oscillator/amplifier. Thesides of the structure, normally in place to form an electrically tight box,have been removed.

13

MIRROR

OPIALPT

AM LFE

-WS

Figure 4b: Photo of N2 laser system showing optical path.

14

MIRROR -----------...............----ob- ---------N2 AMPLIFIER

175 MM LENSA

50 MM LENS

N - ------

N2 OSCILLATORA

HE-NE LASER

BEAM

Figure 5: Optical path of oscillator/amplifier system.

BEAMSPLITTEP

FIBER LASER EEAM

MIRRORVACUUM WINDOW

Figure 6: Mach.Zehnder interferometer.

15

LENS

LASER PLASMA FOCAL PLANE FILMGRATINGS APERTURE

Figure 7: Moirg-Schlieren optics.

Figure 8: Moir-g-Schlieren fringe conditions.

16